RegionServer Failed to Start Because the Port Is Occupied

Symptom

RegionServer is in the Restoring state on Manager.

Cause Analysis

1. View the RegionServer log (/var/log/Bigdata/hbase/rs/hbase-omm-xxx.log).
2. Run the lsof -i:21302 command (the port number of MRS 1.7.X and later versions is 16020) to view the PID. Based on the PID, check the process. It is found that the RegionServer port is occupied by DFSZkFailoverController.
3. The value of /proc/sys/net/ipv4/ip_local_port_range is 9000 65500. The temporary port range and the MRS port range overlap. This is because the preinstall operation is not performed during installation.

Solution

1. Run the kill -9 DFSZkFailoverController pid command to ensure that another port is bound with after a restart and restart the RegionServer in the Restoring state.